Skip to content

Dataset · Joschka

BIG-Bench Hard dataset for LLM evaluation

Joschka/big_bench_hard

BIG-Bench Hard data adapted for integration into the Inspect AI evaluation framework.

What's inside

Includes English text data for question-answering, multiple-choice, and text-generation tasks. Dataset size ranges from 1K to 10K examples.

What you can do with it

  • Evaluate language model capabilities using the Inspect AI framework
  • Benchmarking LLMs on complex logical tasks
Tasks
Question answering, Multiple choice, Text generation
Languages
English
Size
1K–10K rows
Downloads this month
12.4K
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("Joschka/big_bench_hard")